Instructions for running MultiCamSelfCal inside docker container.

This is necessary because MultiCamSelfCal does not appear to work in Octave version > 4.

If docker is not already installed on your system, please install using the official Docker documentation.

1. Clone the repository

In the terminal, navigate to where you plan to install this repository. Clone this repo onto your Desktop with the following:

git clone https://github.com/vanbreugel-lab/multicamselfcal_camera_calibration_in_docker.git

2. Build the Docker image

1. Enter the repository directory: cd multicamselfcal_camera_calibration_in_docker

2. For convenience, save the full path of the repository location as an environment variable in your .bashrc: echo 'export MCSC_PATH=~/multicamselfcal_camera_calibration_in_docker' >> ~/.bashrc

3. Navigate to the docker_MultiCamSelfCal folder: cd $MCSC_PATH/docker_MultiCamSelfCal

4. Build a Docker image from within the docker_MultiCamSelfCal folder:

sudo docker build -t multicamselfcal_image .

3. Run Interactive Container and External Mount

Run the Docker image with the following tags to start the Docker container with an interactive bash shell and live data mount:

sudo docker run --name multicamselfcal_env -it -v ~/Desktop/braid_camera_calibration/docker_mount:/home/host_mount multicamselfcal_image bash

You should see the command prompt change to something like this: root@bf2fb5040035:/home#

This command puts the shared directory in ~/Desktop/braid_camera_calibration/docker_mount. You can change that if you want it somewhere else. The Docker container can interact with the host system through that mount, and vice-versa. Put whatever files you want to use in ~/Desktop/braid_camera_calibration/docker_mount, and you will see it in the /home/host_mount directory of the docker container. Similarly, any files written to /home/host_mount will be visible in ~/Desktop/braid_camera_calibration/docker_mount

4. Test with sample data

In a different terminal (not the one running the docker):

Copy the sample data into the docker_mount: sudo cp -r $MCSC_PATH/example_data/sample_calibration_data.braidz.h5.recal ~/Desktop/braid_camera_calibration/docker_mount

Now the sample data will be accessible to the docker container.

5. Run the MultiCamSelfCal command

In the docker container:

The Docker container should have the following structure:

home
|-- host_mount
`-- MultiCamSelfCal

Navigate to: cd /home/MultiCamSelfCal/MultiCamSelfCal and run:

octave gocal.m --config='/home/host_mount/sample_calibration_data.braidz.h5.recal/multicamselfcal.cfg' 

This will take a while, but it should converge.

Exiting, killing, and restarting the docker container

Helpful commands:

• From inside the docker container, ctrl-d will exit, but leave the container in an inactive state.
• To list all docker containers, including inactive ones: sudo docker ps -a
• To close a docker container: sudo docker rm CONTAINER_ID where CONTAINER_ID is the ID you see when you run the command above.
• To restart a inactive container: docker start container_name